This seems pretty basic, but I can’t figure out how to do it.
Is there a way, given the name of a registered image, to find out what components it’s part of?
Hi Aaron
Best if do a bit “inspection” on how it all work, then take a look on this again https://www.youtube.com/watch?v=d8naLEtLqDY
and start experimenting on smaller datasets to get it in hand, so get not lost from beginning on big data…
Am I misunderstanding, or you did you just not answer my question?
I understand your video and have a good feel for your workflow. My job is big, and the project is much harder than something like a house. It’s 1.5km of cave passage. You could watch this video I did after using Photoscan to put it together last year: https://youtu.be/hz2ZRb-2tXg . I have returned to the site and added 500 more photos since then to fill in the gaps and thought I would try RC on it.
One of the main differences between a project like this and your house is that it is not possible to look at an image and figure out what part of the scene it’s from, because the cave walls look pretty similar the whole way through.
It’s coming along ok – I have one component with 938 images aligned, one with 222, and I am gradually incorporating the other thousand.
Here’s a screenshot to give you a feel.
Anyway, it’s a simple question – is it possible, from the 2Ds view or 1Ds image list, to find out what component(s) a particular image is in? I am guessing the answer is no, it is not possible, which is ok, just need confirmation so I stop trying to figure out how to do it.
If you select the image in the list it will be highlighted in the model and vice versa. If you only have a few components you view each component to find it. Bigger camera icons makes it easier.
So the answer is no, there’s no way to get RC to tell you what components an image is in. It sure would save me a lot of time if this information was added to the “Selected input(s)” panel.
Hi Aaron Curtis
Can you make screenshot of the ALIGNMENT settings used here ?
Sure. I’ve been changing settings, but here’s the part I’m working on right now. I’ve been compromising accuracy just to try and get everything to fit together. I even ran one with feature reprojection error=20 and I think I get a lot of stray points outside of the cave when I do that. I was thinking that once I get it all together I’ll be able to re-run it with tighter settings. Does that make sense?
Ben Kreunen wrote:
If you select the image in the list it will be highlighted in the model and vice versa. If you only have a few components you view each component to find it. Bigger camera icons makes it easier.
Another alternative is to select the image, change the panel mode to 1Ds so it lists your components and then expand the components to list their images. When you select an image it is highlighted everywhere the image is referenced… 3D view, components image list, CP image list